Genome sizes of animal RNA viruses reflect phylogenetic constraints

Animal genomes are characterized by extensive variation in size. RNA viruses similarly exhibit substantial genomic diversity, with genome lengths ranging from 1.7 to ∼64 kb. Despite the myriad of novel viruses discovered by metagenomics, we know little of the factors that shape the evolution of the genome size in RNA viruses. We analyzed the variation in genome sizes across orders and families of animal RNA viruses. We found that RNA viruses can have highly variable genome sizes within and among orders, with the Nidovirales (including the Coronaviridae) exhibiting both significantly larger genomes and a greater range of genome sizes than other orders. In the Bunyavirales, Amarillovirales, Nidovirales, and Picornavirales, the genome sizes of invertebrate-associated RNA viruses were significantly larger than those that infect vertebrates, in contrast to their animal hosts in which vertebrates commonly have larger genomes than invertebrates. However, in the Mononegavirales, vertebrate viruses were significantly larger than those viruses associated with invertebrates. There were similarly complex associations between genome size and patterns of genome segmentation. In the Bunyavirales, Reovirales, and Nidovirales, viruses with segmented genomes, or that possessed a large number of segments, had significantly larger genome sizes than viruses with nonsegmented genomes or a small number of segments, while in Articulavirales, there were no significant differences in genome size among viruses possessing any number of genome segments. More broadly, our analysis revealed that taxonomic position (i.e. RNA virus order) had a greater impact on genome size than whether viruses infected vertebrates or invertebrates or their pattern of genome segmentation. Hence, the phylogenetic constraints on genome size are of sufficient magnitude to shape some other aspects of virus evolution.